Documentation · v1

Competitive intelligence for clinical trials.

Ogur watches clinical trial registries, FDA filings, patent offices, company IR feeds, and the scientific literature — detects competitive changes — and synthesizes them into structured briefings for pharma strategy teams. This site is the technical documentation; for the source code, browse the repository.

01 — Start here

First contact.

• Installation

  ---

  Prereqs, environment variables, seeding the SQLite DB, frontend bootstrap.

  docs/installation

• Architecture

  ---

  System diagram, data model, engine internals, KIQ + verification gate, dedup invariants.

  docs/architecture

• Contributing

  ---

  Make-target catalogue, code style, recipes for adding sources / agents / analyzers.

  CONTRIBUTING

02 — Reference

What lives where.

• API reference

  ---

  Every FastAPI endpoint with request/response schema, query parameters, status codes, curl examples.

  docs/api-reference

• Data sources

  ---

  Per-source authentication, rate limits, signal types produced, upstream documentation links.

  docs/data-sources

• Frontend

  ---

  Vite + React stack, view structure, the Inspector pattern, structured-entity-ref rendering.

  docs/frontend

• Testing

  ---

  Fixture catalogue, the patch-where-used rule, factory functions, the StaticPool pattern.

  docs/testing

03 — Quality

How we measure.

• Evaluation

  ---

  Briefing-quality axes, the entity / outcomes eval harnesses, KIQ-shape verification gate, cost-per-run budget.

  docs/evals

• Research

  ---

  Provenance-first design, agent architecture, BIOPSY benchmark, retrieval seam, competitive positioning.

  docs/research

04 — Design

The visual contract.

• UX specification

  ---

  Canonical visual + interaction contract. The file that wins when implementation disagrees.

  docs/design/ux-spec

• Agentic architecture notes

  ---

  Coordinator/specialist pattern inspired by Claude Code, applied to the synthesis pipeline.

  docs/design/agentic-architecture-notes

Project status

Where Ogur is right now.

Phase	Status	Scope
Phase 1 — Ingestion	Complete	10+ sources, content-hash dedup, ~1,500 signals per landscape run
Phase 2 — Intelligence engine	Complete	Detect → classify (5 domain agents) → enrich → synthesize → verify; QueryEngine for Q&A
Phase 3 lite — REST API	Complete	6+ FastAPI endpoints incl. comparative evidence cards
Harness layer — KIQs + verification gate	Complete	Per-landscape KIQs as DB rows; deterministic gate between synthesis and persistence
Evidence layer — Phase A/B	Complete	Entity NER (GLiNER + Claude), outcomes extraction (Haiku tool-use), CT.gov v2 parser, two eval harnesses
Phase 4 — Frontend	In progress	Vite/React, 4 views, structured entity-ref rendering, KIQ answers section
Phase 3 full — Watcher, semantic dedup, vector search	Planned	APScheduler, Signal Bus, Voyage AI embeddings